Our company’s cost-per-PR dropped from $80 to $30 by switching to GPT 5.6 Sol.

We built a benchmark to replay our team's agent runs across model providers. GPT 5.6 Sol got the highest code quality at a 66% lower cost vs. our old default (Claude Opus 5).

WarpSep 98

Opus 5 made a choir of faces sing this good morning song

w̸͕͂͂a̷͔̗͐t̴̙͗e̵̬̔̕r̴̰̓̊m̵͙͖̓̽a̵̢̗̓͒r̸̲̽ķ̷͔́͝Sep 974

Opus 5 built me a guitar for $28.5.

$18.8 of that was re-reading its own context. One prompt, 75 minutes, 2,142 lines. Zero assets — wood, bark, grass, all in code. You don't pay for output. You pay rent on 240K of context, 157 times over. It watched itself: 53 screenshots +

Andrei ProvkinSep 91

Instead of having Claude tokens idle, I want to make a series

Redesign of public sites in Indonesia. #1 mrtjakarta This redesign uses models (Fable 5.1 + Opus 5) Before:

DiptaSep 91